Signal-Based Selling Comes To Broadcast TV When the Ad Server Meets The Traffic System

Broadcast television generates tens of billions in annual ad revenue, but the way most of that inventory gets priced has barely changed in half a century. Stations sell on age demographics using average quarter-hour ratings as the unit of trade. The system works well enough when the only competition is other broadcasters running the same playbook. It falls apart when the buyer's alternative is a streaming platform that can tell them exactly which households visited a competitor's website this week.

Tom Sly spent more than a decade as VP of Enterprise Strategy at The E.W. Scripps Company, where he built the company's programmatic advertising business into a unit generating over $125 million a year. Now Managing Partner at the consultancy Media Inno, he has deep operational experience across connected TV, FAST channels, and ATSC 3.0. As he recently argued in TVNewsCheck, dynamic ad insertion does not weaken broadcast. It expands what broadcasters can offer by moving them from a single-product model built on mass reach to a multi-dimensional one that combines reach, precision targeting, and measurable outcomes.

He believes the core problem is an infrastructure gap that prevents stations from acting on signals they either already have access to or could acquire through partnerships with TV OEMs and data providers. "The ultimate signal is deterministic," says Sly. "Zero-party or first-party data. I know exactly who they are and I know this specific information about them."

The streaming platforms and TV manufacturers already operate on this logic. Amazon, which functions as the second- or third-largest search engine, can see when a consumer has been researching a specific product category and serve ads across media accordingly. The Walmart acquisition of Vizio extends that same model to the living room screen. Broadcast, for the most part, is locked out of this ecosystem.

The Same Impression, Four Times the Price

The disconnect becomes clearest at the point of sale. Sly describes a scenario in which a broadcaster sells a linear impression to a local Chevy dealer at a $22 CPM because the station has no way to differentiate that impression from any other. The same impression, viewed by the same household, might be worth $80 to Mercedes-Benz if the ad server knew that 5,000 of the 100,000 viewers had visited Mercedes-Benz.com that week.

"When you join an ad server and a traffic system together and use AI for optimization, you get the opportunity to ask whether that $22 impression is actually the best deal," Sly explains.

The technology to make that determination exists. The two systems that would need to talk to each other—the ad server that handles targeting and the traffic system that manages scheduling and reconciliation—are simply not connected in most broadcast operations. That separation is the structural reason broadcast television cannot participate in the kind of signal-driven pricing that digital platforms take for granted.

Why the Plumbing Broke the First Time

The industry had a working proof of concept roughly seven years ago. Project OAR, a consortium led by Vizio with members including Disney, NBCUniversal, and CBS, used watermarks and Ad-IDs to enable dynamic ad insertion on linear feeds. The technology performed as designed. "It worked perfectly," Sly says. "The problem was the traffic systems couldn't handle reconciliation."

When a dynamic ad replaces a scheduled spot, the traffic system needs to account for what was replaced, what ran in its place, and how the revenue allocation shifts. Without that debit-and-credit framework, buyers cannot trust the delivery reports and sellers cannot manage their inventory accurately. The industry moved on because the economics did not force the issue. "Everything was rosy. Nobody pushed for it."

Counting Ads at the Zip Code Level

Sly is now working with a former Scripps consultant on renewed infrastructure tests that pick up where OAR left off. Using watermarks and Ad-IDs, the team can count ad impressions at the zip-code level and show exactly what ran where. In a Salt Lake City test bed, the system maps impression delivery down to individual zip codes: 950 impressions in Fruitland, Utah; 5,493 in Evanston, Wyoming; 28,135 in Cedar Valley.

That granularity is the foundation for everything else. Audience-based selling, yield optimization, dynamic pricing, and eventually real-time AI-driven impression allocation all depend on a trusted counting layer.

The next phase requires collaboration with TV OEMs to enable watermark reading at the device level. That collaboration is complicated by the legal environment. Texas recently sued five smart TV manufacturers over how they collect automatic content recognition data, and California has pursued similar claims. Sly views those legal pressures as making watermark-based approaches more attractive, not less, because they offer a path to signal resolution that sidesteps the ACR mechanisms under scrutiny.

The Reach-Precision Hybrid

The conventional framing positions broadcast reach and digital precision as competing strategies. Sly argues they become complementary once DAI is enabled on linear feeds.

"At any one time, between 10% and 12% of consumers are in the market for an automobile," Sly says. "Why would you want to reach 100% of the audience when what you're looking for are those in-market consumers?"

For upper-funnel campaigns, broadcast reach remains unmatched. For lower-funnel precision buys, broadcasters currently lose the budget to streaming because they cannot deliver targeting or post-campaign measurement, and that streaming ecosystem is scaling fast. A hybrid model, where the same broadcaster sells broad reach on the linear feed and precision impressions through a DAI-enabled stream under one buy, lets the station compete across the entire funnel.

The Consolidation Trap

Several major broadcaster groups are focused on M&A and deregulation as their primary strategic path. Sly sees a risk in that emphasis. "If you consolidate, what does that mean? More stations with the exact same problems you have today?" he says. "As an industry, shouldn't we be focused on truly transforming?"

Acquiring more stations without modernizing the underlying ad-tech infrastructure simply scales existing limitations. The make-good cycle, where under-delivered weight gets manually reconciled through back-and-forth between buyers and sellers, persists whether a group owns 20 stations or 200. Meanwhile, the large agency holding companies are building AI-powered buying platforms that will increasingly expect automated, data-driven transactions from their media partners.

Start With the Consumer

Sly's prescription for broadcaster leadership is to reverse the typical strategic planning sequence. Rather than starting with what the station can sell and how consolidation might expand that footprint, start with what the viewer is experiencing.

"What are they experiencing, and how are they entering their TV experience?" Sly asks. "And then, how do the advertisers want to engage with them? That's what we should be focused on."

The viewer who sits down in front of a connected TV in 2026 enters through the TV manufacturer's electronic program guide, not the cable company's. The first screen they see, the recommendations they receive, and the data trail they generate are all mediated by the OEM. For broadcasters without a collaborative relationship at that layer, the consumer experience is something that happens around them, not through them.

The path forward is straightforward in concept and demanding in execution: connect ad servers to traffic systems, enable watermark-based counting for reconciliation, partner with TV OEMs for device-level signal resolution, and layer AI optimization on top.

"It works. It's proven. It's a matter of commitment," Sly concludes.